Joseph J Amon et al

RESEARCH COMMUNICATION

Trends in Liver Cancer, Province

Joseph J Amon1,2, Supalert Nedsuwan3, Somrak Chantra4, Beth P Bell1, Scott F Dowell5, Sonja J Olsen5, Annemarie Wasley1

Abstract

Objective: To determine liver cancer trends in , Thailand. Methods: Death certificate (1993- 2003) and hospital records (1999-2003) were reviewed and compared to national averages and other provinces. Results: According to data from death certificates, liver cancer mortality in Sa Kaeo Province increased from 3.1 to 26.1 per 100,000 population between 1993 and 2003. In Thailand overall rates increased from 9.0 to 19.8 per 100,000 population between 1996 and 2003. According to electronic hospital records, the total number of patient encounters (in-patient admissions and out-patient visits) for liver cancer in the two main hospitals in Sa Kaeo Province increased 56% (14% annually) between 1999 and 2003. The number of cases of hepatocellular carcinoma increased from 42 in 2001 to 73 in 2003, while the number of cases of cholangiocarcinoma showed little change. Conclusions: Thailand as a whole and Sa Kaeo Province specifically have a high burden of liver cancer, which appears to have increased substantially in the past 10 years. Demonstrating the impact of ongoing strategies aimed at reducing risk factors for liver cancer, such as universal hepatitis B vaccination of infants, will require reliable data describing liver cancer disease burden and etiology. Rapid investigations using available data from death certificates, electronic admissions records, and patient charts can provide valuable insights on disease burden and trends.

Key Words: liver cancer - trends - epidemiology - hepatitis - Thailand

Asian Pacific J Cancer Prev, 6, 382-386

Introduction

Worldwide, liver cancer is the fifth most common cancer including hepatitis B virus (HBV) and hepatitis C virus (Parkin et al., 2001). However, rates of liver cancer and the (HCV) infections and exposure to aflatoxin. In contrast, the relative frequency of the two main types (hepatocellular main risk factor for cholangiocarcinoma is high intensity carcinoma and cholangiocarcinoma) vary significantly infection with the liver flukes Opisthorchis viverrini or between countries, as well as in different regions of the same Clonorchis sinensis. Despite the large burden of disease country (Terry, 2003). Hepatocellular carcinoma, which associated with liver cancer, surprisingly limited information comprises approximately 85% of malignant hepatic tumors is available on trends in the incidence of hepatocellular worldwide, has a high incidence in Sub-Saharan Africa and carcinoma or cholangiocarcinoma globally, or in Thailand. China, and an intermediate incidence in other parts of Asia, Sa Kaeo Province, located in along the Latin America and Europe (Bosch et al., 1999). border with , has an overall population of Cholangiocarcinoma is less common worldwide than approximately 600,000 (2003). In 2002, the single leading hepatocellular carcinoma; however, in specific regions of cause of death reported on death certificates in Sa Kaeo Southeast Asia, most notably the northeast Thailand province Province was cancer. Among cancer deaths, liver cancer was of Khon Kaen, cholangiocarcinoma represents a majority the leading type. In 2003, anecdotal reports from medical of liver cancer cases (Haswell-Elkins, et al., 1994; personnel in Sa Kaeo Province suggesting that liver cancer Srivatanakul et al., 2004). Known risk factors for cases had been sharply increasing over the past few years hepatocellular carcinoma and cholangiocarcinoma differ, prompted an evaluation of liver cancer cases and liver cancer with major risk factors for hepatocellular carcinoma trends in the province.

1Division of Viral Hepatitis, Centers for Disease Control and Prevention, Atlanta, Georgia; 2Epidemiologic Intelligence Service, Epidemiology Program Office, Centers for Disease Control and Prevention, Atlanta, Georgia; 3Bureau of Epidemiology, Thai Ministry of Public Health, Nonthaburi, Thailand; 4Crown Prince Hospital, Sa Kaeo, Thailand; 5International Emerging Infections Program, Thai MOPH – U.S. CDC Collaboration, Nonthaburi, Thailand. Correspondence to: Joseph Amon E-mail: [email protected] 382 Asian Pacific Journal of Cancer Prevention, Vol 6, 2005 Liver Cancer Trends in Sao Kaeo Province, Thailand Methods summarized using a standardized two-page data abstraction form. The information collected included: age, sex, residence To describe general trends in mortality attributed to liver (including village, sub-district, district and province), cancer, death certificate records were reviewed for the period ethnicity, smoking and alcohol use, clinical characteristics, 1996 – 2003 for Thailand as a whole, as well as Sa Kaeo laboratory results (including blood, stool, and ultrasound, if Province and four other provinces from different regions present), and referral information. Based on the final within the country. To further assess trends and describe the diagnosis (as specified on the last medical visit), each patient specific characteristics of liver cancer cases in Sa Kaeo was classified as: 1) hepatocellular carcinoma; 2) Province, hospital records for liver cancer patients diagnosed cholangiocarcinoma; 3) liver cancer, unspecified; or 4) other. between 1999 and 2003 were reviewed. Data were analyzed using SPSS 11.0 (Chicago, IL) to examine trends over time and frequency of different case Death Certificate Review characteristics. Pearson’s chi-square was used for measuring Death certificates for the period 1993-2003 were difference in frequencies, and ANOVA was used for evaluated by review of reports published by the Ministry of comparing median ages. Interior. A case was defined as a death attributed to primary liver cancer, as indicated by an ICD10 code of C22 on the Results certificate of death. Crude mortality rates were calculated using provincial and national census estimates. Rates of liver Death Certificate cancer were compared to those for other tumors (breast, The analysis of death certificate data indicated that, ICD10 code C50; cervix, ICD10 C53; trachea, bronchus, within Thailand as a whole, there has been a sharp increase lung, C33-34; colon, rectum, anus, ICD10 C18-21) for which in mortality attributed to liver cancer, with rates per 100,000 ultrasound is typically used in diagnosis. population increasing from 9.0 to 19.8 between 1996 and 2003. A similar increase was seen across five provinces, with Hospital Record Review in reporting a greater Sa Kaeo Province has a total of 8 hospitals, with in- than four-fold increase (Table 1). In Sa Kaeo Province, patient capacity ranging from 30-230 beds. However, the two main hospitals, (150 beds) and Crown 40 Prince (230 beds) Hospitals, in Eastern and Western regions of the province respectively, have the most advanced staffing 35 and diagnostic capacity (including the only ultrasound equipment), and combined see 75% of in-patients and 60% 30 of out-patients in the province. Interviews with provincial 25 health officers indicated that all suspect cases of liver cancer would most likely be first seen at, or referred to, 20 Aranyaprathet or Crown Prince Hospital. Therefore, the Rate per 100,000 15 assessment of admissions and patient records was limited to these two hospitals. 10 In-patient and out-patient electronic records with an ICD10 diagnostic coding of C22 (primary liver cancer) from 5

1999-2003 for the two hospitals in Sa Kaeo Province were 0 reviewed. A case was defined as an in-patient or out-patient 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 with a diagnosis of primary liver cancer and an ICD10 coding Year of C22. Cases were attributed to the year that a diagnosis of Figure 1. Cancer Mortality (rate per 100,000 population) liver cancer was first noted. All cases were required to be by Type, Sa Kaeo Province, Thailand. (Legend: Solid residents of Sa Kaeo Province. markers: squares = liver; triangles=trachea, bronchus, lung, Patient charts were available for the period 2001-2003. diamonds=colon, rectum, anus. Clear markers: triangles=breast; Data in the medical charts of individual patients were diamonds=cervix)

Table 1. Liver Cancer Rate per 100,000 Population, 1996 – 2003, in Province 1996 1997 1998 1999 2000 2001 2002 2003 Sa Kaeo 12.5 12.2 10.5 6.5 15.8 26.8 34.2 26.1 Khon Kaen 13.2 12.5 21.4 22.3 26.2 34.6 41.4 35.8 Nakhon Phanom 14.8 14.4 22.4 23.8 27.4 29.4 31.8 54.3 Chiang Rai 9.4 10.7 2.7 3.1 8.6 9.0 12.8 19.5 Ranong 2.1 1.4 1.3 1.3 4.4 9.3 4.3 9.8 National Average 9.0 9.5 12.0 12.8 14.6 16.5 17.7 19.8

Asian Pacific Journal of Cancer Prevention, Vol 6, 2005 383 Joseph J Amon et al greater increases were seen in liver cancer than cancers of carcinoma with serological testing results for HBV or HCV the trachea/bronchus/lung, breast, cervix or colon/rectum/ markers, 46% (6 of 13) were hepatitis B surface antigen anus (Figure 1). (HBsAg) positive and 40% (2 of 5) were antibody to hepatitis B core antigen (anti-HBc) positive. Among individuals Hospital Records diagnosed with cholangiocarcinoma, 13% (1 of 8) were According to electronic hospital admissions records, the HBsAg positive and 0% (0 of 2) were anti-HBc positive. total number of patient encounters (in-patient and out- Thirteen percent (1 of 8) of hepatocellular carcinoma patients patient) for liver cancer in Crown Prince and Aranyaprathet had antibody to HCV (anti-HCV) compared to 0% (0 of 1) Hospitals increased 56% (14% annually), from 174 to 271 of cholangiocarcinoma patients. Stool examination was also cases, between 1999 and 2003 (Table 2). The increase was done on a limited number of patients. Among individuals somewhat greater at Crown Prince Hospital (15% annually), diagnosed with hepatocellular carcinoma, 7% (1 of 15) had which also reported 69% of the overall number of cases over a positive result on stool examination (for Strongyloidies the time period, than at Aranyaprathet Hospital (12% sp.), while the only positive result among the stool annually). Little difference was seen in liver cancer examinations (1 of 13, or 8%) among cholangiocarcinoma admission rates according to district of residence within the patients was due to the presence of Opisthorchis eggs. Of province (data not shown). the limited number of biopsy reports available, most (11 of Patient medical charts were available for 95% of the 719 14) were inconclusive. patients who were identified by electronic record review as having been diagnosed with liver cancer during the years Discussion 2001-2003 (Table 2). Twenty-three percent (23%) of the records located were found to have a diagnosis other than Liver cancer is an important public health problem in liver cancer and had been miscoded as C22; these were many parts of the world, but details of disease burden, excluded from further analysis. temporal trends, and the factors underlying these trends are Among the 521 patients diagnosed with liver cancer not well documented. Using death certificate data and during 2001-2003, the number of cases attributed to hospital records we conducted a rapid examination of trends hepatocellular carcinoma increased from 42 in 2001 to 73 in liver cancer rates in Sa Kaeo Province, Thailand, and in 2003. During the same period, the number of individuals found evidence that one type of liver cancer in particular, diagnosed with cholangiocarcinoma remained hepatocellular carcinoma, was increasing. The investigation approximately constant, accounting for 40 cases in 2001 and demonstrated that available data sources can provide 44 cases in 2003. The number of cases with a diagnosis of ‘ valuable information on liver cancer trends and underlying liver cancer, unspecified’ or ‘other’ also remained etiologies. Improving the quality and availability of this type approximately constant (Table 2). of data, and conducting periodic follow-up analyses can Comparing the characteristics of patients classified with further increase our understanding of the burden and hepatocellular carcinoma (n=167) and cholangiocarcinoma epidemiology of liver cancer in high disease areas, and help (n=143), patients with a diagnosis of hepatocellular assess the impact of primary prevention strategies. carcinoma were slightly younger (median age of 55 vs. 60 years, p=.05), more likely to be male (78% vs. 62%, p=.005), Improving our Understanding of Liver Cancer Trends and and more likely to frequently consume alcohol (19% vs. Causes 10%, p=.02). The two groups of patients had a similar In Thailand, death certificate data indicated that mortality prevalence of abdominal pain, hepatomegaly, and abdominal from reported liver cancer has increased between 1996 and mass. 2003, and in Sa Kaeo Province increases were seen since Limited laboratory data were available to suggest 1993. Similar increases were not seen for other major cancer etiologies. Among individuals diagnosed with hepatocellular types. Hospital records from Sa Kaeo Province provided

Table 2. Trends in Diagnosis of Liver Cancer, by Hospital and Specific Type, Sa Kaeo Province, Thailand, 1999 - 2003 Year Variable 1999 2000 2001 2002 2003 # Liver Cancer Admissions (from electronic hospital records) 174 206 205 243 271 # (%) Patient Records Found 196 (96%) 229 (94%) 255 (94%) # (%) Patient Records Eligible 148 (76%) 187 (82%) 186 (73%) Patient Classification Hepatocellular carcinoma 42 (28%) 52 (28%) 73 (40%) Cholangiocarcinoma 40 (27%) 59 (32%) 44 (24%) Liver Cancer, unspecified 59 (40%) 63 (34%) 58 (31%) Other 7 (5%) 10 (5%) 10 (5%)

384 Asian Pacific Journal of Cancer Prevention, Vol 6, 2005 Liver Cancer Trends in Sao Kaeo Province, Thailand additional evidence for an increase in the number of liver just less than one percent (0.9%) of blood donors were anti- cancer cases, however, this increase was substantially smaller HCV positive. Information on aflatoxin exposure, another than death records indicated. important cause of hepatocellular carcinoma, is not available The specific nature of the increase and the underlying in Sa Kaeo Province, although a number of studies conducted factors are unclear. Primary liver cancer is difficult to in other parts of Thailand have failed to find an important diagnose, and limited information on case-patients was role for aflatoxin exposure in local cases of hepatocellular available – few patients were referred for liver biopsy, and carcinoma (Srivatanakul et al., 1991a; Srivatanakul et al., few had complete laboratory information. While changes in 1991b; Wild et al., 1990). diagnostic equipment, staffing or protocols, or in patient Although there does not appear to be a sharply increasing utilization of health care services or demographics might trend in cases of cholangiocarcinoma in Sa Kaeo Province, theoretically account for the observed trends, there was no the finding of 24%-32% of cases being due to evidence that these changes occurred in Sa Kaeo Province. cholangiocarcinoma is higher than expected for this region The absence of similar increases for other types of cancer in of Thailand, given the low frequency of O. viverrini infection Sa Kaeo Province also suggests that the increase in hospital reported. Limited information on trends of Opisthorchis admissions for liver cancer represents a true increase in infection in Sa Kaeo Province, and limited diagnostic testing primary liver cancer incidence. of liver cancer patients for the presence of O. viverrini The term liver cancer encompasses a heterogeneous infection, make it difficult to understand possible trends in group of specific diagnoses reflecting different etiologic risk factors associated with cholangiocarcinoma, and to factors. Improvements in diagnostic accuracy could help to estimate the possible future burden of cholangiocarcinoma focus prevention strategies on relevant risk factors. While in the province. definitive diagnosis of liver cancer usually relies on In addition to improving the availability and use of histologic examination, in high-incidence regions of the specific diagnostic tools to more accurately classify liver world, the differentiation of hepatocellular carcinoma and cancer as hepatocellular or cholangiocarcinoma, efforts to cholangiocarcinoma can be made with a high degree of improve the quality of death certificate reports and hospital accuracy based on a combination of imaging and laboratory record coding can greatly improve the ability to conduct measures (Kew, 1997). For example, a high or rising serum rapid assessments of liver cancer of trends. Specific studies alpha-fetoprotein (AFP) level (greater than 400 ng/ml), and to assess diagnostic accuracy (for example using post- a hypervascular hepatic mass (>2 cm) on radiographic mortem autopsy) and the establishment of cancer registries imaging studies is characteristic of hepatocellular carcinoma can provide further, and more detailed, information to (Bruix et al., 2001). In contrast, patients with complement death certificate and hospital record-based data. cholangiocarcinoma typically have normal serum AFP levels, elevated serum alkaline phosphatase levels, and more Implementing Effective Strategies for Primary Prevention frequently present with jaundice (Ahrendt et al., 2001; Zhou of Liver Cancer et al., 2000). In these patients, ultrasound typically shows Despite uncertainties about the trend and underlying dilated intrahepatic ducts and normal extrahepatic ducts. causes of liver cancer in Sa Kaeo Province, prevention Despite limitations in accurately distinguishing measures are needed to address the undisputedly high rates hepatocellular carcinoma from cholangiocarcinoma in our of liver disease. Because treatments options are limited, and data, it appears that the majority of the increase in liver cancer median survival short, primary prevention strategies, that mortality can be attributed to a rise in hepatocellular is, prevention aimed at eliminating or reducing exposure to carcinoma. HBV infection is the most frequent underlying those factors that cause liver cancer, represent the best cause of hepatocellular carcinoma worldwide, and in opportunity to alter the natural history of liver cancer. Thailand between 8-10% of males and 6-8% of females are In Sa Kaeo Province, the majority of liver cancer cases estimated to be HBsAg positive (Merican et al., 2000). HBV were due to hepatocellular carcinoma, and the majority of prevalence in Sa Kaeo appears to be similar: among antenatal these cases were likely due to HBV infection. Hepatitis B clinic attendees at one hospital in 2002, 6% were HBsAg vaccination is the most effective means of HBV infection positive. However, without more detailed information on prevention. Because the virus is often transmitted from trends in HBV prevalence, or on age-specific HBV mother to child, and the risk of chronic infection is greatest prevalence, understanding the underlying dynamics driving for newborns and young children, vaccination at an early increasing rates of hepatocellular carcinoma changes is age is necessary. Hepatitis B vaccination has been integrated difficult. Consequently, it is also difficult to predict whether into Thailand’s expanded program on immunization since the current increase will continue, and if so, for how long. 1992, and the effect of vaccination on reducing chronic HBV In addition to HBV infection, other factors may also be infection has already been demonstrated (Poovorawan et al., contributing to the high burden of hepatocellular carcinoma 2000). In order for Thailand to demonstrate the impact of in Sa Kaeo Province, although most likely to a smaller extent. reduced infection on liver cancer rates (something Taiwan, HCV infection is another underlying cause of hepatocellular with twenty years of vaccination efforts has demonstrated carcinoma, and is found in between 1.5% and 3% of adults among children (Chang et al., 1997; Chang, 2003; Huang et in Thailand (Nantachit et al., 2003). In Sa Kaeo Province, al., 2000)), improved data and repeated assessments are Asian Pacific Journal of Cancer Prevention, Vol 6, 2005 385 Joseph J Amon et al needed. Chang MH, Chen CJ, Lai MS, et al (1997). Universal hepatitis B However, recognizing that current newborn vaccination vaccination in Taiwan and the incidence of hepatocellular campaigns in Thailand will take decades to impact upon carcinoma in children. Taiwan Childhood Hepatoma Study liver cancer rates, other primary prevention measures Group. N Engl J Med, 336, 1855-9. directed at older persons should not be forgotten. Both HBV Haswell-Elkins MR, Mairiang E, Mairiang P, et al (1997) Clinical and nonimaging diagnosis of hepatocellular carcinoma. in: and HCV transmission can be reduced by ensuring effective Liver cancer, eds: Okuda K, Tabor E. Churchill Livingstone, blood screening and proper sterilization and safe needle use. New York, 315-359 Because of the lack of a hepatitis C vaccine, this approach Huang K, Lin S (2000). Nationwide vaccination: a success story is particularly important in preventing HCV infection. in Taiwan. Vaccine, 18, S35-8. In addition to targeting hepatocellular carcinoma, Larrey D (1997). Hepatotoxicity of herbal remedies. J Hepatol, 26 primary prevention strategies can address causes of Suppl 1, 47-51. cholangiocarcinoma. To reduce the burden of Loapaiboon V, Sithithaworn P, Elkins DB (1994). Cross-sectional cholangiocarcinoma, public health education campaigns can study of Opisthorchis viverrini infection and spread messages to reduce the consumption of raw fish, and cholangiocarcinoma in communities within a high-risk area in northeast Thailand. Int J Cancer, 59, 505-9. sanitation efforts can be conducted to reduce fecal Merican I, Guan R, Amarapuka D, et al (2000). Chronic hepatitis contamination of water sources. B virus infection in Asian countries. J Gastroenterol Hepatol, Although the extent to which alcohol and other hepatoxic 15, 1356-61. exposures (such as herbal remedies) influence the burden Nantachit N, Robison V, Wongthanee A, et al (2003). Temporal of liver cancer in Sa Kaeo Province is uncertain, primary trends in the prevalence of HIV and other transfusion- prevention campaigns can also include education to limit transmissible infections among blood donors in northern these factors, which can have an additive impact on Thailand, 1990 through 2001. Transfusion, 43, 730-5. hepatocarcinogenesis (Larrey, 1997) . Parkin DM, Bray FI, Devesa SS (2001). Cancer burden in the year 2000. The global picture. Eur J Cancer, 37 Suppl 8, S4-66. Poovorawan Y, Theamboonlers A, Vimolket T, et al (2000). Impact Conclusion of hepatitis B immunisation as part of the EPI. Vaccine, 22, 943-9. Liver cancer has increased substantially in Sa Kaeo Srivatanakul P, Parkin DM, Jiang YZ, et al (1991a). The role of Province, Thailand, in the past 10 years. Both hepatocellular infection by Opisthorchis viverrini, hepatitis B virus, and carcinoma and cholangiocarcinoma were important causes aflatoxin exposure in the etiology of liver cancer in Thailand. of liver cancer, however our data indicated that the increase A correlation study. Cancer, 68, 2411-7. was most likely due to increases in the number of cases of Srivatanakul P, Parkin DM, Khlat M, et al (1991b). Liver cancer in hepatocellular carcinoma. Using readily available data, it Thailand. II. A case-control study of hepatocellular carcinoma. was possible to demonstrate both a high burden of disease Int J Cancer, 48, 329-32. Srivatanakul P, Sriplung H, Deerasamee S (2004). Epidemiology due to liver cancer in Sa Kaeo Province, and an increasing of liver cancer: an overview. Asian Pac J Cancer Prev, 5, 118- trend. Given the high rates of HBV infection prevalence in 25. other , similar increases may be occurring Terry WD (2003). Primary cancer of the liver. In: Cancer. Ed: elsewhere. However, this analysis also shows the limitations Kaplan HS. New York: Alan R. Liss, 101. inherent in documenting and understanding the dynamics Wild CP, Jiang YZ, Allen SJ, et al (1990). Aflatoxin-albumin underlying liver cancer burden in high prevalence areas of adducts in human sera from different regions of the world. the world given the limited available data. Improving our Carcinogenesis, 11, 2271-4. understanding of the burden of disease and changing trends Zhou XD, Tang ZY, Yang BH, Zhang BH (2000). Hepatocellular in liver cancer in areas of high prevalence can help in the Carcinoma: The Role of Screening. Asian Pac J Cancer Prev, 1, 121-6. design of effective prevention efforts, and provide evidence of their value and impact.

References

Ahrendt SA, Nakeeb A, Pitt HA (2001). Cholangiocarcinoma. Clin Liver Dis, 5, 191-218. Bosch FX, Ribes J, Borras J (1999). Epidemiology of primary liver cancer. Semin Liver Dis, 19, 271-85. Bruix J, Sherman M, Llovet JM, et al (2001) EASL Panel of Experts on hepatocellular carcinoma. Clinical management of hepatocellular carcinoma. Conclusions of the Barcelona-2000 EASL conference. European Association for the Study of the Liver. J Hepatol, 35, 421-30. Chang MH (2003). Decreasing incidence of hepatocellular carcinoma among children following universal hepatitis B immunization. Liver Int, 23, 309-14. 386 Asian Pacific Journal of Cancer Prevention, Vol 6, 2005